Preferences for this copy of Plexora.
Every project you import, and every figure you build, is written here. Plexora reads this once when it starts, so a change takes effect the next time it is launched.
{# Filled by settingsPage.js from /settings/data. Rendered empty rather than server-side because the same markup has to be redrawn after a save, and one renderer beats a Jinja copy and a JS copy that drift. #}Existing projects stay where they are unless you choose to bring them along on the next step.
Plexora runs on one machine and reaches every other one over SSH.
{# The other half of that sentence, and the half people come here looking for. The note above says the panel tells them where the launch decision is made; saying it takes the command, because "start it there" with no example is not an instruction anybody can act on. #}
To run Plexora itself on one of them, start it there
rather than from here —
plexora connect you@login.cluster.edu in a
terminal, or Open OnDemand, or a Jupyter session. From that
Plexora the cluster is Local, and these same saved servers
are what plexora connect reads.
Save a remote connection once, then reconnect by name.
Start from the kind of machine you use. Plexora fills in what it knows; you can change any of it afterwards. If your institution's cluster is one Plexora already knows, it is under Additional presets.
Machines that are handing this Plexora data right now. Most
of them arrive by themselves: choosing Remote
on a data field opens one on a saved server, and
plexora connect opens one on your own computer.
Nothing here needs setting up before you import.
Which files a project reads from a node is chosen on that project’s own page. This is only the address book.
For a node you started yourself with
plexora node serve on a machine this one can
already reach. A node opened from a data field or by
plexora connect registers itself.
An image opened straight from a web address is read through a cache on this disk, one chunk at a time, so the same bytes are not fetched twice. This is where that cache is, how big it may get, and what is in it right now.
Kept inside the data directory. Moving the data directory in the Data section above takes this cache with it, and clearing it never touches the images themselves, only this copy of their bytes.
need an
extra package: pip install 'plexora[remote]'
Options for private buckets or S3-compatible servers, saved once per web address prefix. Keys are never stored here — Plexora uses your AWS/Google/Azure credentials from the environment.
Plexora looks for a newer release once a day and puts a dot on the Help menu when there is one. It never installs anything until you ask.
Plexora can send anonymous usage counts — which features are used, how long tiles take to draw, which kinds of errors occur — so we know what to improve. It never sends file names, project names, marker or gene names, cell data, coordinates, usernames, machine names or anything typed into a tool. Everything is queued on this machine first, and nothing in Plexora waits for it or depends on it.
Also off with plexora telemetry off or
DO_NOT_TRACK=1.
Plexora Free is everything Plexora does without a licence: every image, table and modality, every tool's manual features, remote and HPC viewing, notebooks and export — with no account and no network. Paid adds AI features: guided gating sessions and the evidence an agent gathers, with more to come.
For a computer with no internet access. Run
plexora license fingerprint --out fp.json there,
upload it in the licence portal, then choose or paste the
.plexora file here.
Your data never leaves this machine: a licence check sends the certificate and a hash that identifies this environment, nothing else. When a licence ends, everything Free keeps working and everything you made with Paid features stays yours.